Primary human nasal epithelial cells: a source of poly (I:C) LMW-induced IL-6 production

Infection plays a significant role in the relapse of chronic rhinosinusitis (CRS), however, the role of primary human nasal epithelial cells (HNECs) in this process is largely unknown. Here, we determined the effect of Toll-like receptor (TLR) agonists and inflammatory cytokines on mucosal barrier integrity and immune response of HNECs. TLR 1–9 agonists and inflammatory cytokines were applied to submerged and/or air-liquid interface (ALI) cultures of HNECs from CRS patients and controls for 24 hours. Interleukin-6 (IL-6) protein levels were determined by ELISA. Mucosal barrier integrity was measured via Transepithelial Electrical Resistance and passage of fluorescently-labelled dextrans. IL-1β and IFN- γ significantly increased IL-6 production in HNECs derived from CRS patients and controls, however, a dose-dependent effect was observed in CRS-derived HNECs only. Stimulation with Poly (I:C) LMW induced a 15 to 17 fold increase in IL-6 production by HNEC-ALI control cells (p < 0.05) and HNEC-ALI-CRS cells (p = 0.004) whilst a 2.5 fold increase was observed in CRS HNEC submerged cultures. Priming of cells with Poly (I:C) LMW reduced subsequent IL-6 secretion upon stimulation with TLR 2–4 agonists. Poly (I:C) LMW exerts a potent pro-inflammatory effect on HNECs and reduces a subsequent immune activation by TLR agonists.

The sinonasal mucosa has been widely recognised as protecting the host from invasion by harmful environmental toxins and micro-organisms by forming a structural barrier. The epithelial apical junctional complex (AJC) which comprises tight and adherens junctions, is critical to maintain mucosal barrier integrity and epithelial cell polarity. Disruption of AJC proteins leads to mucosal barrier dysfunction and is frequently found in severe chronic inflammatory diseases of the gut, skin and airway 1,2 . The role of the airway mucosa in raising and shaping an immune response to different environmental insults has been extensively described and different model systems have been developed. These include ex vivo mucosal explant models that have been shown to have a robust response to bacterial triggers 3,4 . The advantage of such models is that they adequately mimic the in vivo situation as they represent the combined immune response of the different immune cell types present within the mucosa to such triggers. The disadvantage is that such explant models are inherently stressed due to lack of adequate oxygen and nutrient supply within the tissue, and are viable only for a limited amount of time (depending on the challenge from 24-72 hours) 3 . Also, the mucosa comprises a range of different cell types known to be critical in orchestrating such responses and a specific role of airway epithelial cells within that process has not been fully elucidated 5,6 . Airway epithelial cell culture models are widely used, as such cells are easy to grow and give consistent results with relatively low variability between experiments. However, airway epithelial cell lines will, in general, not form a functional barrier structure and mucociliary transport system and they do not have a conserved innate immune response machinery, hence any findings on the immune response when such cells are used should be interpreted with caution 7 . Primary human nasal epithelial cells (HNECs) are equipped with innate immune receptors and can respond to a range of environmental insults of microbial 7 and non-microbial origin 8 , contributing to the immune response to those triggers. HNECs cultured at ALI can differentiate into a ciliated, pseudostratified epithelium that secretes mucus, exerts high Trans Epithelial Electrical Resistance (TEER) (a measure of the epithelial barrier function) and has a functional mucociliary transport system, mimicking the Air Liquid Interface Culture. HNECs were maintained at Air Liquid Interface (ALI), following the Lonza ALI culture method (Lonza, Walkersville, USA) as described previously 11,13 . Briefly, 7 × 10 4 HNECs were seeded in a volume of 100 µL B-ALI medium which contains (Bovine Pituitary Extract [BPE], Hydrocortisone, human Epidermal Growth Factor [hEGF], Epinephrine, Transferrin, Insulin, Retinoic Acid, Triiodothyronine, Gentamicin/Amphotericin-B, Bovine Serum Albumin -Fatty Acid Free [BSA-FAF] and inducer) into the apical chamber of Transwell plates (BD Biosciences, San Jose, California, USA) and 500 µL of B-ALI growth medium was added to the basal chamber in all wells and incubated for 3-4 days at 37 °C with 5% CO 2 . Then, the apical media was removed and 500 µL B-ALI ™ differentiation medium was added to the basal chamber. The cultures were fed every other day by adding B-ALI complete differentiation medium to the basal chamber. HNECs at air liquid interface (HNEC-ALI) were maintained for a minimum of 21 days for development of tight junctions.
Enzyme-Linked Immunosorbent Assay (ELISA). Supernatant was collected from the basolateral compartment of treated HNEC-ALI cultures after 24 hours of exposure with inflammatory cytokines. Interleukin-6 (IL-6) protein levels were estimated with an ELISA kit using rat anti-human IL-6 antibodies (BD Biosciences, New Jersey, USA), according to the manufacturer's instructions. All measurements were performed in duplicate using a FLUOstar OPTIMA plate reader (BMG Labtech, Ortenberg, Germany). The tissue sample concentration was calculated from a standard curve and corrected for protein concentration.
Transepithelial Electrical Resistance (TEER). Transepithelial electrical resistance (TEER) was measured by using an EVOM volt-ohmmeter (World Precision Instruments, Sarasota, FL, USA). Briefly, 100 µL of B-ALI medium was added to the apical chamber of ALI cultures to form an electrical circuit across the cell monolayer and into the basal chamber. Cultures were maintained at 37 °C during the measurement period using a heating platform. Only wells displaying baseline resistance readings greater than 700 Ω/cm 2 were used for the experiments. TLR agonists and control (B-ALI medium for the negative control and 2% Triton × 100 for the positive control) were added to the basal and/or apical chambers of each Transwell and TEER measurements were Immunofluorescence microscopy. Cells were fixed with 2.5% formalin in phosphate-buffered saline (PBS) for 10 min. Fixed samples were permeabilized with 0.1% Triton X-100 in PBS for 15 minutes, blocked for 1 hour with Protein Block (Dako), and incubated with 2 μg/ml rabbit polyconoclonal anti-human TLR3 (#LS-B4866, Sigma, Aldrich, USA), overnight at 4°C. In negative controls, the primary antibody was replaced with PBS. Excess primary antibody was removed, and 2 μg/ml anti-Rabbit CY3 conjugated secondary antibody (Jackson ImmunoResearch Labs Inc., West Grove, PA, USA) was added and incubated for 1 hour at RT. The Membranes were rinsed in TBST, and after the third wash, 200 ng/ml of 4′, 6-diamidino-2-phenylindole (DAPI; Sigma, Aldrich, USA) was added to resolve nuclei. Membranes were transferred to a glass slide and a drop of anti-fade mounting medium (Dako, Glostrup, Denmark) was added before cover-slipping. Samples were visualized by using a LSM700 confocal laser scanning microscope (Zeiss Microscopy, Germany). Slide tissue was prepared as above except tissue were cut in 4 µm sections from CRSwNP patients, deparaffinized and rehydrated. Antigen retrieval was induced at 100 °C for 10 minutes in 10 mmol/L sodium citrate buffer, pH 6.
Statistical analysis. Data are presented as mean ± SEM. The IL-6 assays where HNECs were grown in submerged cultures were analysed using t-tests and all other analysis was performed using ANOVA, followed by Tukey's HSD post hoc test using SPSS (version 22). A P value less that 0.05 was considered statistically significant.

Poly (I:C) LMW increased the secretion of IL-6 by HNEC-ALI cultures when applied to both apical and basal Transwell chambers.
Poly (I:C) HMW and Poly (I:C) LMW were added to the apical, basal and both apical and basal chambers of Transwells. Poly (I:C) LMW increased the production of IL-6 to 325 pg/ ml, 629 pg/ml and 3957 pg/ml when applied in apical, basal and both apical and basal chambers respectively compared with negative control (220.9 pg/ml). The protein level of IL-6 was significantly higher when Poly (I:C) LMW was applied in both basal and apical chambers (p = 0.002). In contrast, Poly (I:C) HMW when applied in apical  (322 pg/ml), basal (308 pg/ml) and both apical and basal chambers (435 pg/ml) did not induce any significant increase in the production of IL-6 compared with negative control (220 pg/ml, p > 0.05) (Fig. 4).

TLR agonists did not affect the permeability of HNEC-ALI monolayers. HNEC-ALI cultures were
stimulated with TLR1-9 agonists for 24 hours followed by assessment of the permeability of the monolayers by measuring the TEER and passage of FITC-dextrans at 24 h. Sodium Dodecyl Sulphate (SDS) and B-ALI complete medium were used as a positive and negative control respectively. None of the TLR agonists affected TEER measures (Fig. 5A) or permeability of FITC-dextrans (Fig. 5B).

Priming of HNECs with Poly (I: C) LMW reduces IL-6 secretion after challenge with different TLR agonists.
To determine whether priming of HNEC-ALI cultures influenced subsequent immune responses, Poly (I:C) LMW was added for 24 h to the basal and apical chambers of HNEC-ALI cultures from control patients. The supernatants were removed and fresh media was added for recovering the cells. After 24 h, the primed cells were treated with the TLR agonists HKLM, LPS or Poly (I:C) LMW. IL-6 secretion was measured in  TLR3 is localised to the sinus epithelium and expressed by HNECs. Immunofluorescence of sinus mucosa, HNEC-ALI and HNEC submerged cultures using TLR3-specific antibodies showed TLR3 expression in the epithelium layer in sinonasal mucosa and in HNECs in the nucleus, cytoplasm and cell periphery (Fig. 7).

Discussion
Interleukin-6 is known to be produced by airway epithelial cells in response to microbial stimulation and plays a significant role in chronic allergic airway inflammation 3,4,14,15 . This study showed that HNECs derived from CRSwNP patients had a significantly higher IL-6 production compared to HNECs derived from non-CRS control patients in response to a range of immune stimuli including IFN-γ, IL-1β and Poly (I:C) LMW. In addition, IFN-γ and IL-1β increased IL-6 production in a dose-dependent manner in CRSwNP derived HNECs but not in non-CRS control derived HNECs. Previous studies have similarly shown HNECs derived from CRSwNP patients respond differently to topical treatments from HNECs derived from control patients 12 . These donor-dependent differential responses could indicate the presence of genetic alterations and/or epigenetic modifications in CRS patient-derived HNEC cultures that remain active after multiple cell divisions have occurred in vitro. Interestingly, recent reports indicate the presence of epigenetic modifications in tissue samples from CRSwNP patients 16,17 , even though the relevance of these findings to CRSwNP-derived HNECs is unclear. Similarly, genetic studies have shown polymorphisms in genes involved in antigen presentation, innate and adaptive immune responses, tissue remodeling and arachidonic acid metabolism in association with CRS (reviewed in 18 ). Whether or not (epi) genetic modifications are present in CRSwNP-derived HNECs and whether such modifications might translate to differential innate immune responses of CRSwNP-derived HNECs compared to control-derived HNECs is unknown. Further experiments designed to specifically address these questions are needed to test these hypotheses.
Our results indicate that Poly (I:C) LMW rather than Poly (I:C) HMW consistently induced innate immune activation and IL-6 secretion by HNECs. Both LMW and HMW Poly (I:C) are viral dsRNA surrogates corresponding to viral dsRNA of different lengths (between 0.2-1.0 and 1.5-8 kilobase pairs, for LMW and HMW Poly (I:C) respectively). LMW and HMW Poly (I:C) signal through TLR3 ligation, however, the resulting immune response can differ, with HMW Poly (I:C) inducing more potent TLR3-dependent IFN-β signaling compared to LMW poly (I:C) 19 . In addition to ligating TLR3, LMW Poly (I:C) can also bind Retinoic acid-inducible gene-I (RIG-I) whereas HMW Poly (I:C) can bind the related cytosolic helicase, myeloid differentiation-associated gene 5 (MDA5) 20 . The specific structural characteristics of the dsRNA molecules that optimally activate these innate immune receptors are reflected in the types of viruses that are recognized by those receptors. Namely, RIG-I is primarily responsible for innate immune recognition of paramyxoviruses, influenza virus and Japanese encephalitis virus, while MDA5 is thought to be critical for the recognition of picornaviruses 20 . TLR3 on the other hand, recognizes dsRNA molecules greater than 40-50 bp in length and mediates the induction of antiviral responses to, for example, rhinovirus, respiratory syncytial virus and influenza virus, frequently responsible for viral respiratory infections 21 . Our data implies that HNECs might be more sensitive to those viruses that are able to activate TLR3 and/or RIG-I. It is well known that severe bacterial lung infections are often preceded by viral infections 22,23 . Similarly, in the upper airways, viral infections and associated mucosal damage frequently precede bacterial superinfection and acute sinusitis 24 . Moreover, it has been shown that upper airway influenza virus infection significantly alters the sinonasal microbiome and increases the bacterial burden in both upper and lower airways, significantly augmenting the susceptibility to bacterial pneumonia 25 . Our data indicate that priming of HNECs with TLR3 agonists reduces subsequent innate immune responses to a range of TLR agonists. Whilst the mechanism of this finding and its significance remain to be defined, the reduced sensitivity of HNECs to TLR ligation after TLR3 agonist priming might be due to induction of interferon responses. Namely, TLR3 agonists as well as viral infections are known to induce potent interferon responses in a range of immune cell types 26 . Such responses are needed to fight the viral infection. However, they are accompanied by a reduced generalized innate immune response to different immune triggers, facilitating bacterial superinfection 27 . It will be interesting to identify the exact virus types that can stimulate innate immune responses in HNECs and identify and compare their activation and signaling pathways in HNECs.
In conclusion, our data show that HNECs are equipped with innate immune defense mechanisms that allow for a potent immune activation upon ligation of Poly (I:C) LMW and that HNECs derived from CRSwNP patients react more vigorously to immune triggers than HNECs derived from non-CRS control patients. Moreover, we have shown that stimulation with Poly (I:C) LMW reduces subsequent immune activation with different TLR agonists. Together, these data indicate that HNECs play an important role in the immune activation and regulation upon viral infection of the upper airway.